Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier.com/locate/biortech

Consequence of replacing nitrogen with carbon dioxide as atmosphere onsuppressing the formation of polycyclic aromatic hydrocarbons in catalyticpyrolysis of sawdust

Yao Hea,c, Si Chena, Junjie Chena, Dongxia Liuc, Xunan Ninga, Jingyong Liua, Tiejun Wangb,⁎

aGuangzhou Key Laboratory of Environmental Catalysis and Pollution Control, Guangdong Key Laboratory of Environmental Catalysis and Health Risk Control, School ofEnvironmental Science and Engineering, Institute of Environmental Health and Pollution Control, Guangdong University of Technology, Guangzhou 510006, Chinab School of Chemical Engineering and Light Industry, Guangdong University of Technology, Guangzhou 510006, Chinac Department of Chemical and Biomolecular Engineering, University of Maryland, College Park, MD 20742, USA

A R T I C L E I N F O

Keywords:Lignocellulosic biomassCatalytic pyrolysisCarbon dioxidePolycyclic aromatic hydrocarbonsZeolite

A B S T R A C T

This study evaluates the effect of replacement of N2 with CO2 as atmosphere in catalytic pyrolysis of wastelignocellulosics with acidic and metal-modified zeolites, respectively, on the 16 EPA priority pollutant polycyclicaromatic hydrocarbons (PAHs) in bio-oils. By coupling solid phase extraction pretreatment with single ionmonitoring detection, it is found that the replacement alleviates PAHs in bio-oil concerning synchronouslyabating the 16 PAHs with low, medium and high molecular weights, and the benzo[a]pyrene equivalent toxicityof bio-oil decreases. Meanwhile, CO2 decreases the content of small oxygenates, e.g. furans, ketones, acids, andincreases phenolics and aromatics affording more stable and valuable bio-oils. Moreover, CO2 enhances carbonconversion efficiency, especially in combination with Fe-modified zeolite, which presents a synergistic effect.This study indicates the practical application of CO2 as an atmosphere in catalytic pyrolysis to improve the bio-oil quality by suppressing PAHs formation and adjusting compound constituent.

1. Introduction

The speeding-up social metabolism toward carbon, triggered bymassive exploitation and utilization of fossil resources, has drawn thepublic acceptance for substituting renewable carbon source for tradi-tional petro-derived fuels and chemicals. Considering the carbon neu-trality, biomass such as lignocellulosics that are widely dispersed andeasily accessed has been highlighted as a cost effective and promisingalternative for mitigating the environmental concern resulted fromgreenhouse gas emission (Dabros et al., 2018; Wang et al., 2019).

Pyrolysis that occurs in the absence oxygen at elevated tempera-tures is an efficient route by which biomass can be converted into liquidfuels and renewably derived chemicals. This technology allows biomassto thermally decompose to smaller fragments and to form up a liquid(bio-oil) with high oxygen content and poor stability (Sharifzadeh et al.,2019). With the addition of catalysts, pyrolysis vapor is upgraded priorto condensation, and consequently a relatively more stable bio-oil withless oxygen content is obtained. This process is usually named as cat-alytic pyrolysis, in which the acidic zeolite such as HZSM-5 (Hoff et al.,2017; Kumar et al., 2019) and its metal modified counterparts (Li et al.,

2017; Li et al., 2016; Mullen and Boateng, 2015) are commonly appliedas the catalyst because of their good capabilities for deoxygenatingpyrolysis vapor and producing aromatic-rich bio-oil.

However, the use of zeolite catalysts is a double-edged sword. Forexample, an increase in valuable aromatics typically comes along withan increase in undesired byproduct, polycyclic aromatic hydrocarbons(PAHs). PAHs are a group of persistent organic contaminants with morethan two aromatic rings, most of which are teratogenic, mutagenic,carcinogenic, etc., and therefore they are extremely harmful for health(Mahler et al., 2012). As a result, the exposure of bio-oil, e.g. during itsstorage, handling, or transportation, would pose potential risks forhuman and ecosystem. For monitoring purposes, the U.S. Environ-mental Protection Agency (EPA) has made a list of 16 unsubstitutedPAHs that are on a priority pollutant list. In addition, PAHs in bio-oilcan also cause major challenges in downstream post-processing reactorsand decrease the overall carbon conversion efficiency of the process(Stark & Ghoniem, 2017). Nevertheless, most literatures concerning thetopic of PAHs produced from biomass pyrolysis focused on the PAHsadhered to solid residues or particulate matters released (Dandajehet al., 2018; Ko et al., 2018; Konczak et al., 2019), and fewer efforts

https://doi.org/10.1016/j.biortech.2019.122417Received 6 October 2019; Received in revised form 7 November 2019; Accepted 9 November 2019

⁎ Corresponding author.E-mail address: tjwang@gdut.edu.cn (T. Wang).

Bioresource Technology 297 (2020) 122417

Available online 12 November 20190960-8524/ © 2019 Elsevier Ltd. All rights reserved.

T

http://www.sciencedirect.com/science/journal/09608524https://www.elsevier.com/locate/biortechhttps://doi.org/10.1016/j.biortech.2019.122417https://doi.org/10.1016/j.biortech.2019.122417mailto:tjwang@gdut.edu.cnhttps://doi.org/10.1016/j.biortech.2019.122417http://crossmark.crossref.org/dialog/?doi=10.1016/j.biortech.2019.122417&domain=pdf

have been devoted to analyze these PAHs that bio-oil contained. Zhouet al. (2014a) conducted the pyrolysis of xylan, cellulose and lignin at agasification temperature (800 °C), and found the existence of ten spe-cies of PAH from the EPA priority list in bio-oil, which contained nomore than four aromatic rings. Dunnigan et al. (2017) examined thebio-oil produced at 500 °C, of which the results indicated that therewere only two and three ring EPA PAHs. While, similar findings werereported in the studies on spent coffee grounds (Nguyen et al., 2019).However, these researches merely involved the PAH individuals withlower molecular weights among the 16 EPA priority pollutant PAHs andconcerned the bio-oil from traditional pyrolysis of biomass without thepresence of any catalyst. In fact, it is highly significant to earn an intactcognition for all the 16 PAHs in bio-oil, since most of the PAHs withhigher molecular weights, such as benzo[a]pyrene, which typicallyemerge in trace amount and hardly to be detected, are extremely toxic.

Recently, Kwon and coworkers proposed that by replacing N2 withCO2 as the atmosphere in traditional pyrolysis, the formation of PAHscould be effectively mitigated. They have developed a number of in-vestigations on different feedstocks, e.g. sewage sludge (Kim et al.,2019a), polyvinyl chloride (Lee et al., 2018), municipal solid waste (Leeet al., 2017a), printed circuit board (Lee et al., 2017b), styrene buta-diene rubber (Kwon et al., 2012), etc., confirming the suppression im-pact of CO2 atmosphere on PAHs formation. Similarly, the thermaldecomposition of lignin, a typical component of lignocellulosic bio-mass, at a gasification temperature (800 °C) under N2 and CO2 atmo-spheres, also revealed that the content of PAHs that existed in bio-oilproduced under CO2 atmosphere was relatively lower in comparisonwith under N2 atmosphere (Zhou et al., 2014b). Apart from the effecton the PAHs suppression, it has been also found the presence of CO2 canattract a few variations in bio-oil composition. The pyrolysis of corncobunder CO2 atmosphere was likely to provide a more stable bio-oil thatcontained less methoxy groups with respect to under N2 atmosphere(Zhang et al., 2011). Mante et al. (2012) further corroborated thosefindings by using 13C nuclear magnetic resonance technique. They ad-ditionally found that the presence of CO2 atmosphere led to increaseddeoxygenation reactions, which resulted in the reduction of the relativecontent of not only methoxyls, but also sugars, carboxylics and carbo-nyls. In addition, through the studies on the pyrolysis of medicinal herbresidue, Zhang and Zhang (2017) concluded that the use of CO2 as theatmosphere could also arise the hydrocarbon content in bio-oil.

In the light of these previous studies, we assumed that by sub-stituting CO2 for traditional inert gas N2 as the atmosphere, catalyticpyrolysis over zeolites commonly conducted at a lower temperature(~500 °C) can also give rise to a bio-oil with limited content of PAHpollutants and, accordingly, experiments were performed to explore theimpacts of this substitution. The obtained bio-oil was handled withsolid phase extraction pretreatment followed by gas chromatographypaired with spectrometer (GC–MS) analysis at single ion monitoring(SIM) detection mode to avoid strong interference during detectingprocess resulted from organic compounds such as phenolics (Fabbriet al., 2010). This analysis method as a result ensured the determinationof all the 16 EPA priority pollutant PAHs in bio-oil, especially for thespecies with high molecular weights that mostly exist at trace levels butwith high toxicities. In this way, the variation of the 16 PAHs in bio-oilwith the substitution of CO2 for N2 as the atmosphere in catalyticpyrolysis is first time fully revealed, as well as the benzo[a]pyrene

equivalence toxicity of the bio-oil. At the same time, the constituent ofeffective chemical compound that existed in bio-oil produced underCO2 in reference to N2 was also compared. For fundamental study, alab-scale pyrolysis of Fraxinus mandschurica sawdust, a commonlyavailable waste lignocellulosics from furniture manufacturing industryin China, under N2 or CO2 atmosphere was carried out. The widely usedzeolites HZSM-5 and its transition metal modifications were applied asthe catalysts in this study.

2. Materials and method

2.1. Materials and catalyst preparation

Fraxinus mandschurica sawdust was collected from a furniturefactory in Panyu district, Gurangzhou city, China. Commercial proto-nated ZSM-5 zeolite (HZSM-5) from Nankai University Catalyst withSiO2/Al2O3 ratio of 25 was calcined at 550 °C for 4 h to eliminate or-ganic compound and moisture content before use. Transition metal Feand Co modified HZSM-5 zeolite catalysts was prepared by incipientimpregnation using Fe(NO3)3 and Co(NO3)2, respectively, achieving10 wt% metal loading in each case, and followed by drying at 70 °Covernight. The solids were subsequently calcined at 550 °C in air for4 h, and then reduced at the same temperature by N2/H2 (volume ratio9:1) for 2 h. Proximate analysis including moisture (M), ash, volatilematter (VM), and fixed carbon (FC) of the samples, was performedaccording to the American Society for Testing and Materials (ASTM)D1762-84. Ultimate analysis, including carbon, hydrogen, and ni-trogen, was carried out using an elemental analyzer (Vario EL cube,Elementar). The hemicellulose and cellulose components of Fraxinusmandschurica sawdust were quantitatively determined by the methodof ASTM D5896-96, and the quantity of the lignin component wasidentified by ASTM D1106. The results for these basic analyses on thebiomass feedstock are shown in Table 1.

2.2. Catalyst characterization

The crystal structure of the catalysts was characterized via X-raydiffraction (XRD) analysis by Cu Kα radiation at a scan rate of 6°/minwith a step of 0.02° in the 2θ angle ranging from 5° to 50° performed onBruker D8 Advance X-ray diffratremeter. The nature of the acid sites oncatalyst was studied by Fourier transform infrared spectroscopy (FTIR)of adsorbed pyridine on a PerkinElmer Frontier FTIR spectrometer. Thecatalyst sample was first finely grounded in an agate mortar andpressed into a ca. 20 mg self-supporting wafer (13 mm of diameter) inthe absence of binder, which was then placed in an in situ infrared cellwith CaF2 windows. Before the measurements, all samples were treatedunder vacuum (10−2 Pa) at 400 °C for 2 h and then cooled to roomtemperature. Spectra of the pretreated samples were collected asbackground. Next, pyridine was introduced to the in situ IR cell wherethe samples were saturated for 1 h at room temperature. The samplewas then evacuated for 10 min to remove the unabsorbed pyridine. Thedesorption of the pyridine was successively monitored stepwise, byevacuating the sample for 30 min at 350 °C, and cooling to roomtemperature to record the spectrum. The difference spectrum was ob-tained by subtracting the background spectrum that was previouslyobtained. Textural properties of the catalyst involving specific surface

Table 1Basic analysis of Fraxinus mandschurica.

Proximate analysis (wt.%) Ultimate analysis (wt.%) Component analysis (wt.%)

M Ash VM FCa C H Oa N S Cellulose Lignin Hemicellulose

0.60 0.92 83.17 15.31 46.03 6.07 46.88 0.03 0.99 40.63 30.20 29.17

a Acquired by subtraction.

Y. He, et al. Bioresource Technology 297 (2020) 122417

2

area, and pore size and volume of the catalysts were determined byMicromeritics ASAP 2020 Plus HD88 analyzer. The nitrogen adsorp-tion/desorption isotherms were conducted at −196 °C. Brunauer-Emmett-Teller (BET) method and Barret-Joyer-Halenda (BJH) methodwere adopted for analyzing the specific surface area and pore size andvolume, respectively.

2.3. Pyrolysis experiment

The catalytic pyrolysis was performed in a tubular reactor at 500 °C.In each run, ~ 4 g of Fraxinus mandschurica sawdust and catalyst orquartz sand mixture at the ratio of 1:1 in weight was loaded in a quartzboat which was then pushed into the heating zone of the tubular reactorwhen the set temperature was reached. A constant N2 or CO2 with300 mL/min was introduced during the pyrolysis to create a relativelyinert atmosphere and carry volatile products out. A catalytic pyrolysisexperiment using HZSM-5, Fe/HZSM-5, and Co/HZSM-5 zeolites underN2 atmosphere is denoted as N2-HZ, N2-Fe/HZ, and N2-Co/HZ, re-spectively; similarly, the use of them under CO2 atmosphere is denotedas CO2-HZ, CO2-Fe/HZ, and CO2-Co/HZ, respectively. A set of methanolabsorption bottle was placed in an ice bath to condense organic sub-stances. After 5 min reaction, the quartz boat was pushed out from theheating zone and the solid residues (bio-char) were cooled down toroom temperature under inert atmosphere. Then, the methanol solu-tions of bio-oil were collected and saved in a refrigerator at 3 °C foranalysis.

2.4. Bio-oil pretreatment and analysis

The bio-oil solutions was undergone several pretreatment proce-dures for the analysis of the 16 EPA priority pollutant PAHs. The so-lution was first dried by anhydrous sodium sulfate overnight. 5 speciesof deuterated PAHs (AccuStandard, USA), i.e, naphthalene-D8, ae-naphthene-D10, phenanthrene-D10, chresene-D12, and perylene-D12,were spiked into the solution as recovery indicators for the 16 PAHswith different vapor pressures (Table 2) to evaluate their loss caused bythe subsequent treatments. The sample was then rinsed by the solutionof n-hexane and dichloromethane mixture in the volume ratio of 7:3,and purified by flowing through a silica gel and alumina packedcolumn. Next, the eluate collected at the bottom of the column was

concentrated to ~1 mL by rotary evaporation and N2 blowing eva-poration, which was then the bio-oil sample adapted to PAHs analysis.Before the analysis of PAHs, hexamethylbenzene (Ehrenstorfer-schaferBgm-Schlosser Laboratory, Germany) was added as the internal stan-dard.

The 16 PAHs were analyzed by gas chromatography (Agilent7890B) paired with mass spectrometry (Agilent 5977B) (GC–MS) usingselected ion monitoring (SIM) mode. The primary and secondary se-lected ions for the 16 PAHs, the 5 deuterated PAHs (dPAHs), and theinternal standard hexamethylbenzene where determined according toEPA method 8270E (SW-846). In this study, HP-5MS(30 m × 0.25 mm × 0.25 μm, Agilent) capillary column was used forseparation purpose in GC–MS analysis. The temperature program ofcapillary column oven for the PAHs analysis was held at 80 °C for 1 minfirst; then increased to 160 °C at 3 °C/min; increased to 220 °C at 5 °C/min and, finally increased to 310 °C at 10 °C/min and held for 5 min.Quantitative analysis of the 16 PAHs was conducted against a calibra-tion curve generated by injecting a series of mixed PAHs standard so-lutions (Ehrenstorfer-schafer Bgm-Schlosser Laboratory, Germany) ofwhich the concentration ranged from 200 to 1000 μg/L. Three dupli-cate test were performed, and the recoveries indicated by the 5 dPAHsranging from 75% to 110% for all tested samples with relative standarddeviation in the range of 4% to 11.5%.

Besides, the bio-oil solution after drying was directly used for semi-quantitative analysis of chemical composition via GC–MS. The MS wasset to be at scanning mode in the range of 30–500 amu. While, thetemperature program of the oven was held at 35 °C for 8 min first; thenincreased to 100 °C at 5 °C/min and held for 3 min; increased to 220 °Cat 5 °C/min, and finally increased to 290 °C at 10 °C/min and held for5 min. The results were acquired by integrating the relative area ofpeaks appeared at different retention times according to the total ionchromatogram.

3. Results and discussion

3.1. Catalyst characterization

To check the potential structure changes of the zeolite frameworkafter the metal modification, XRD analysis was carried out. The relatedXRD patterns corroborate the preservation of ZSM-5 zeolite structure inmetal modified samples, consistent with the earlier reports (Li et al.,2017; Li et al., 2016). The XRD peaks of metal crystals are hardly seenin these patterns. The absence of the peaks of Fe or Co metal materialsuggests that these species should be well-dispersed in the zeolitesupport. On the other hand, a deviation is observed with respect to theintensity of the featuring peaks of ZSM-5: the peaks become smallerafter metal addition, which is possibly due to the decrease in the pro-portion of ZSM-5 crystalline phases in a metal modified zeolite. Li et al.(2017) found a similar result in their research on HZSM-5 zeolite cat-alyst modified by several metals.

Pyridine (Py) has been proved to be an appropriate probe moleculeto test the acidic characteristics of zeolite solids. The merit of its use isthat the molecule adsorbed on Brønsted and Lewis acid sites affords IRbands at different wavenumbers, and thus it renders the differentiationof these sites. However, it is noteworthy that the Py-IR results can onlygive semi-quantitive information towards the variation in amount forboth sites rather than precise quantification. The Py-IR spectrum wasrecorded at 350 °C after weakly adsorbed pyridine was removed byevacuation at the same temperature. The band at 1545 cm−1 and1450 cm−1 corresponds to the pyridine adsorption on Brønsted andLewis acid sites, respectively. The numerical values of integrated in-tensities (Aint) are listed in Table 3. It is found that the HZSM-5 zeoliteexhibits Brønsted acidity only. However, the nature of the aciditychanges with the loading of the transition metal Fe or Co. Lewis aciditystarts to exist in the Fe modified HZSM-5 zeolite, with the Brønsted/Lewis (B/L) ratio of 3.51, and this ratio sharply declines to 0.34 for the

Table 2Characteristics of priority pollutant PAHs regulated by EPA.

PAH compound Abbreviation No. ofrings

Vapor pressurea (Pa,25 °C)

TEFb

Naphthalene Nap 2 11.14 0.001Acenaphthylene Acp 3 3.87 0.001Acenaphthene Ace 3 3.07 0.001Fluorene Flu 3 1.66 0.001Phenanthrene PhA 3 1.06 × 10-1 0.001Anthracene Ant 3 8.60 × 10-4 0.01Fluoranthene FluA 4 8.61 × 10-4 0.001Pyrene Pyr 4 5.00 × 10-5 0.001Benz[a]anthracene* BaA 4 5.43 × 10-4 0.1Chrysene* Chr 4 4.00 × 10-6 0.01Benzo[b]fluoranthene* BbF 5 5.00 × 10-7 0.1Benzo[k]fluoranthene* BkF 5 5.20 × 10-8 0.1Benzo[a]pyrene* BaP 5 6.00 × 10-8 1Dibenzo[ah]

anthracene*DbA 5 1.33 × 10-8 1

Indeno[123-cd]pyrene* IcP 6 1.27 × 10-7 0.1Benzo[ghi]perylene BgP 6 1.38 × 10-8 0.01

* Carcinogenic PAHs.a According to (Odabasi et al., 2006).b Toxicity equivalent factor (TEF) refers to the toxicity of each PAH species

compared to that of benzo[a]pyrene which is the highest equal to 1 (Nisbet andLagoy, 1992).

Y. He, et al. Bioresource Technology 297 (2020) 122417

3

Co metal modification. As shown in table 3, compared to Fe/HZSM-5,Co/HZSM-5 zeolite is further more Lewis acidic and much less Brønstedacidic, of which the trend is in accordance with the findings for theaddition of Fe and Co on MCM-41 zeolites (Szegedi et al., 2004). At theimpregnation step in a metal modification treatment, metal cations areprone to be anchored at a Brønsted acid site that owns a proton, whichmay accordingly cause the replacement for the proton or the coverageover the Brønsted acid site with the metallic species (Bernardon et al.,2017), and thus causes decrease in Brønsted acidity and increase inLewis acidity.

The results obtained from N2 adsorption/desorption isotherms arelisted in Table 3. It shows that specific surface areas diminish after theaddition of Fe and Co, leading to ca. 7% and 22% decrease, respec-tively. This suggests the appearance of the metal species inside theporous ZSM-5 network (Iliopoulou et al., 2012), which is also reflectedby the decrease in total pore volume. However, the average pore dia-meters are nearly identical for all the samples so that the shape se-lectivity of the channels may be less affected by the metal loadings (Xuet al., 2017).

3.2. Composition of bio-oils

The relative content of different group of chemical compounds inliquid products collected from Fraxinus mandschurica pyrolysis withand without catalysts over zeolites under N2 and CO2 atmosphere isshown in Fig. 1. Six groups of chemical compounds were categorized,including furans, phenols, aromatics, ketones and miscellaneous oxy-genates (mis-oxy). Phenols and furans are the most abundant groups in

the bio-oils, accounting for 18.81–36.06% and 18.19–32.00%, respec-tively. The content of aromatics, ketones, and acids are 1.78–23.84%,13.75–21.30%, and 1.57–6.64%, respectively. Aldehydes and alcoholsare involved into the category of miscellaneous oxygenates, which is inthe range of 6.15 to 12.64%.

Aromatics are most value-added chemical compound within bio-oilfrom catalytic pyrolysis with zeolite, since they can be directly used asliquid fuel additives (Lu et al., 2018a; Lu et al., 2018b; Zhang andZhang, 2017). Apparently, the use of HZSM-5 and Fe/HZSM-5 zeoliteleads to an increase in aromatics, while this effect is less significant forCo/HZSM-5 zeolite. Under N2 atmosphere, the highest content of aro-matics 20.00% is achieved via applying HZSM-5 owing to its strongerBrønsted acidity and higher surface area, followed by Fe/HZSM-5(12.19%), and the content is only 3.05% for the application of Co/HZSM-5, which is comparable to the amount in the bio-oil obtainedfrom the pyrolysis without zeolite. This observation indicates thatHZSM-5 is more effective than are its metal modified counterparts forthe formation of aromatics. The findings from Mullen and Boateng(2015) also presented that the addition of transition metals on HZSM-5would suppress the formation of aromatics, and HZSM-5 had a higherproduction rate than its iron modification when it came to aromatics.By replacing N2 with CO2 as the catalytic pyrolysis atmosphere, the bio-oils are more aromatic. Particularly, the increase in aromatics is mostobvious when using HZSM-5 and Fe/HZSM-5 zeolite catalysts. Theyhave a relative content of 23.84% and 18.01%, which corresponds toca. 1.2 and 1.5 times more than the amount of aromatics observed in N2atmosphere, respectively. Hence, it can be said that CO2 atmosphere isable to enhance the formation of aromatics. This observation is co-herent in the light of studies led by Mante et al. (2012): participation ofCO2 in pyrolysis atmosphere results to the increase in total aromatics.They ascribed this phenomenon to the capability of CO2 for suppressingcoke formation and reducing carbon deposits on the catalyst resultingin its better catalytic performance. Prior studies on conversion of me-thanol to aromatics also corroborated the positive impact of CO2 on thearomatization reaction over zeolite (Xu et al., 2017). They claimed thatzeolite is capable of activating the CO2 absorbed on its surface as ahydrogen receptor to prevent the olefin intermediates from formingparaffins by hydrogenation, thus facilitating the yield of aromatics.

Phenols are an important chemical family for industrial. For ex-ample, they are starting materials to make plastics, explosives such aspicric acid, and drugs such as aspirin. In this study, phenolics are one ofthe most abundant species, which can be possibly resulted from thatphenolics are mainly derived from thermal decomposition of lignincomponent of biomass, of which the content is as high as ~30 wt% inFraxinus mandschurica as shown in Table 1. Under N2 atmosphere, theuse of HZSM-5 zeolite leads to the fewest phenols (18.81%) in bio-oilcompared to the cases with Co/HZSM-5 (21.76%) and Fe/HZSM-5(31.67%) catalyst, and without any catalyst (27.09%). As for the effectof reaction atmosphere, CO2 seems to have enhanced phenols formationsince their content shows an increase in peak area in comparison withN2. These findings are consistent with the literature where phenols havebeen reported to increase in bio-oils from the pyrolysis of switchgrasswith substituting CO2 for N2 as the atmosphere (Pilon & Lavoie, 2013).Besides, similar to the trend in N2 atmosphere, Fe/HZSM-5 exhibits thebest performance for phenols yield under CO2 atmosphere, corre-sponding to the relative peak area of 36.06%, followed by Co/ZSM-5 atthe value of 26.68% that is slightly less than the case without catalyst(31.85%). The worst performance for phenols yield under CO2 atmo-sphere is also presented by HZSM-5 giving a relative peak area of21.96%.

In parallel, the content of furans, ketones, carboxylic acids and otheroxygenates in bio-oil is changed with the application of different cat-alyst and atmosphere applied in pyrolysis as well. For example, the useof HZSM-5 reduces the content of furans, and the addition of Fe on itfurther intensifies this reduction giving a relative peak area of 23.05%that is the lowest yield among the bio-oils obtained under N2

Table 3Textural and acidic characteristics of catalysts.

Catalyst N2 adsorption/desorption Py-IR

SBET (m2/g) Dpore(nm)

Vtotal(mL3/g)

Brønsted Ainta

(cm−1)Lewis Ainta

(cm−1)

HZSM-5 295.93 1.35 0.20 0.28 n.d.Fe/HZSM-5 274.67 1.35 0.18 0.13 0.03Co/HZSM-5 229.81 1.32 0.16 0.03 0.10

a SBET denotes specific area.b Vtotal denotes total pore volume.c Dpore denotes average pore diameter.d Aint denotes integrated intensity of Py-IR signal.

Fig. 1. Content of different chemical species in bio-oils produced under N2 andCO2 atmospheres with HZSM-5, Fe/HZSM-5 and Co/HZSM-5 zeolite catalystand without catalyst.

Y. He, et al. Bioresource Technology 297 (2020) 122417

4

atmosphere, while Co/HZSM-5 gives rise to a similar content of furanswith respect to the case without catalyst. The replacement of N2 withCO2 generally leads to a minor abatement in the content of furans ex-cept for the catalyst of Fe/HZSM-5. Similarly, it can be also observedthat the presence of CO2 mostly gives a lower content of ketones, acids,and miscellaneous oxygenates. It is worth noting that furans, a typicalcategory of intermediates formed via thermal fragmentation of thecellulose and hemicellulose constituents, are basically consumed bysecondary degradations such as dehydration, decarbonylation, dec-arboxylation, oligomerization, and aromatization which are prone tooccur at the Brønsted acid sites on a zeolite (Che et al., 2019). In re-ference to Table 3, it can be concluded that the strong Brønsted acidityof zeolite would enhance furan species conversion, which was alsocorroborated previously (Kumar et al., 2019). In addition, it is observedthat ketones are slightly fewer with the application of HZSM-5, but thentheir content is increased by the use of the metal modified zeolites.Earlier studies has also reported that compared to parent HZSM-5, themodification of transition metals incurred an increase in ketones(Mohabeer et al., 2019).

The above observations signify that the substitution of CO2 resultsin an enhancement for deoxygenation reactions affording a bio-oil withfewer oxygen-containing compounds. These compounds such as acidsand ketones are mostly active compositions in pyrolytic bio-oil(Iliopoulou et al., 2012). Thereby, it can be inferred that to use CO2instead of N2 as the pyrolysis atmosphere is able to make the bio-oilmore stable. At the same time, the resulted increase in its aromatic andphenolic constituents can be associated with a more value-added bio-oil. According to the investigations on bio-oils produced from pyrolysisof hybrid poplar through 13C Nuclear Magnetic Resonance analysis,carboxylic and carbonyl fractions were lessened and aromatic fractionswere increased in the presence of CO2 in pyrolysis atmosphere (Manteet al., 2012), of which the result is aligned with the findings from thecurrent study. These phenomena are plausibly due to the expeditedthermal cracking of oxygenates as a result of employing CO2, which waspreviously evidenced in the pyrolysis of spent coffee grounds (Kimet al., 2019b).

3.3. CO2 suppression for PAHs

3.3.1. PAHs determinationPolycyclic aromatic hydrocarbons (PAHs) are ubiquitous con-

taminants in the environment. It is therefore important to know thecontent of PAHs present in bio-oil before its application. Applying solidphase extraction pretreatment combined with GC–MS single ion mon-itoring to the PAH analysis of bio-oil enables the detection of PAHs thatoccur in trace amount, in particular for PAHs with higher molecularweights often exhibiting high toxicities. The 16 EPA priority pollutantPAHs are detected in the bio-oil obtained from pyrolysis in N2 atmo-sphere in the absence of catalyst, and their distribution and con-centration are presented in Fig. 2. The PAHs are categorized into threegroups: high molecular weight with 5–6 aromatic rings (HPAHs) in-volving BbF, BkF, BaP, DbA, IcP, and BgP, medium molecular weightwith 4 aromatic rings (MPAHs) involving FluA, Pyr, BaA, and Chr, andlow molecular weight with 2–3 aromatic rings (LPAHs) involving Nap,Acp, Ace, Flu, PhA, and Ant. These PAHs are of environmental concerndue to their possible toxicity in humans and other organisms and theirprevalence and persistence in the environment (Dat and Chang, 2017).

The concentration of the PAHs ranges from 0.53 to 5.29 μg/g, wherenaphthalene is the most abundant species while benzo[ghi]perylenepossesses the least amount. This result is mostly in line with the pre-vious findings that the 16 PAHs were in the range of 0.1 to 0.7 μg/g(Fabbri et al., 2010). In general, the content of the PAHs decreases withthe increase in their molecular weight. It is noteworthy that the mea-surement error that shown by the error bar of each column also appearsa decrease as the molecular weight of the PAHs increase, which can beimmediate relevance to the vapor pressure of each species presented in

Table 2. In relative terms, PAHs with high vapor pressure, e.g. naph-thalene, acenaphthylene, acenaphthene, fluorene, which have ad-vanced vaporizing tendency, are less controllable in regards to the lossduring bio-oil pretreatment process, and thus give rise to a relativelyhigh error. Likewise, as the result of the variation in vapor pressure, therecoveries indicated by the deuterium labeled surrogates are ca. 80%for naphthalene-D8 and aenaphthene-D10, which represent the loss ofLPAH species, and close to 100% for phenanthrene-D10, chresene-D12and perylene-D12 representing the loss of PAHs with higher molecularweights and lower vapor pressures. In addition, it is observed fromFig. 2 that the total PAHs of low ring is more than 10 μg/g, but the totalPAHs of medium and high ring is less than half of the total PAHs of lowring. Similar trends were also previously reported by Dunnigan et al.(2017) for the bio-oil collected from pyrolysis of rice husk in N2 at-mosphere. However, their results are in a much smaller quantity(0.85–0.09 μg/g, 4–6 rings not detected). This discrepancy is possiblyattributed to the nature of feedstock, the operating conditions of thereactor, for example, the reactor type and the carrier gas injection rate,and the PAH analysis method as well.

3.3.2. CO2 effect on PAHs with different molecular weightsFig. 3 shows the content of LPAHs, MPAHs, and HPAHs in the bio-

oils produced at different catalytic scenarios under N2 and CO2 atmo-sphere, respectively. LPAHs had a much larger amount compared toMPAHs and HPAHs, giving a variation from ten to hundred, andnaphthalene is always the most abundant PAH species. In contrast, thechange of MPAHs and HPAHs is relatively small. Their concentrationvaries within the range of 1 to 5 μg/g. The predominance of naphtha-lene in its content among the PAH species that the bio-oil contained wasalso evidenced in other studies on the pyrolysis of a variety of biomassfeedstocks such as lignin (Zhou et al., 2014b), pine (Li et al., 2016),palm kernel shell (Maliutina et al., 2017) and spent coffee ground(Nguyen et al., 2019). Among the given 16 EPA PAHs, benzo[a]pyreneis of the greatest concern due to its high carcinogenicity (Wang et al.,2017). Fortunately, the available data showed that benzo[a]pyrene isless available in the obtained bio-oils (< 1 μg/g), lower than the meanvalue (1.3 μg/g) of 48 different crude oils that most commonly apply inreal practice (Kerr et al., 1999).

As for LPAHs presented in Fig. 3(a), a clear increase in their total isobserved with the use of HZSM-5 catalyst, which is over 35 times morethan those detected in bio-oil produced without catalyst. This sig-nificant increase could be mitigated by loading transition metals on thezeolite. For example, the LPAHs concentration decreases from351.10 μg/g when HZSM-5 is applied, to 71.25 μg/g when the Fe-modified zeolite is applied, and it is further decreases to < 20 μg/gwith the application of Co/HZSM-5 zeolite. By substituting CO2 for N2as the pyrolysis atmosphere, it gives rise to a general decrease in LPAHs

Fig. 2. Distribution and quantity of the 16 EPA priority pollutant PAHs in bio-oil produced under N2 atmosphere without catalyst.

Y. He, et al. Bioresource Technology 297 (2020) 122417

5

concentration. As presented in Table 4, the substitution reduces theLPAHs production catalyzed by Fe/HZSM-5 most, arriving at ~35%reduction, which corresponds to the absolute reducing amount of92.33 μg/g. While from Fig. 3(b) and 3(c), it is found that the totalamount of MAPHs and HPAHs in bio-oil is quite low (< 5 μg/g) inreference to LPAHs, which complies with the findings reported byAlhroub et al. (2018) who also claimed that PAH species with highermembered rings is hard to formed for a short thermal treatment time. Insome detail, a decrease of MPAHs and HPAHs is generally observedwith the addition of catalysts, and particularly the use of HZSM-5zeolite reduces their concentrations to the level of< 2 μg/g. Again, thepresence of CO2 brings about a detrimental effect on the both PAHcategories. For example, the substitution of N2 by CO2 as an atmosphereleads to nearly a half reduction in the total amount of MPAHs andHPAHs, respectively, in the absence of catalyst.

As a result, it can be said that the use of CO2 instead of N2 as anatmosphere leads to a synchronous reduction in the concentration ofLPAHs, MPAHs, and HPAHs in bio-oil obtained from the pyrolysis ofFraxinus mandschurica biomass no matter with or without catalyst. Atthe same time, it is worth underlying that two- and three-ringed PAHscan easily dissolved in bio-oil, while PAHs with higher molecularweights are prone to adhere to particular matter such as carbon re-sidues and zeolite solids during pyrolysis (Ramesh et al., 2004). This

could be a part of the reason for the decline in the content of MPAHsand HPAHs in bio-oils with the use of zeolite catalysts.

3.3.3. CO2 effect on toxicity and carbon conversion efficiencyIn Fig. 4(a), the gray and orange columns stand for the sum of the

concentrations of 16 EPA priority pollutant PAHs (∑PAH16) and 7carcinogenic PAHs (∑PAH7), respectively, and the black dots representthe toxic equivalence quotient (TEQ) of the 16 PAHs, i.e. the benzo[a]pyrene equivalence toxicity of bio-oil, where TEQ = ∑(PAH i × TEF i)

Fig. 3. Comparison of LPAHs (a), MPAHs (b), and HPAHs (c) in bio-oils produced under N2 and CO2 atmospheres with HZSM-5, Fe/HZSM-5 and Co/HZSM-5 zeolitecatalyst and without catalyst.

Table 4Reduction effect of CO2 on PAHs in bio-oil.

Parametera CO2 vs. N2 CO2-HZ vs.N2-HZ

CO2-Fe/HZ vs.N2-Fe/HZ

CO2-Co/HZvs.N2-Co/HZ

LPAHs-D (μg/g) 2.68 92.33 24.16 2.56LPAHs-P 26.59% 26.30% 33.92% 13.18%MPAHs-D (μg/g) 1.84 0.57 0.63 1.26MPAHs-P 42.67% 29.37% 19.11% 35.17%HPAHs-D (μg/g) 1.93 0.50 0.40 1.16HPAHs-P 43.22% 28.19% 12.88% 31.21%∑PAH16-D (μg/

g)6.45 93.39 25.19 4.98

∑PAH16-P 34.20% 26.32% 32.45% 18.63%∑PAH7-D (μg/g) 2.48 0.64 0.54 1.58∑PAH7-P 42.91% 28.07% 13.72% 32.76%TEQ-D 0.71 0.32 0.22 0.56TEQ-P 37.49% 26.79% 14.18% 31.44%

a LPAHs-D, MAPHs-D, HPAHs-D, ∑PAH16-D, ∑PAH7-D, and TEQ-D are re-duction amounts equal to the value in CO2 minus the value in N2, respectively;LPAHs-P, MPAHs-P, HPAHs-P, ∑PAH16-P, ∑PAH7-P, and TEQ-P are reductionpercentages equal to the value of reduction amount divided by the value in N2.

Fig. 4. ∑PAH16, ∑PAH7 and TEQ of the 16 EPA priority pollutant PAHs (a);ratio of aromatics against the sum of EPA 16 priority pollutant PAHs (b).

Y. He, et al. Bioresource Technology 297 (2020) 122417

6

in which PAH i and TEF i denotes the concentration and TEF of acertain PAH species, respectively. Obviously, the addition of zeolitecatalysts causes a significant increase in ∑PAH16 of which the amountfollowed the order HZSM-5 > Fe/ZSM-5 > Co/ZSM-5 under both N2and CO2 atmospheres. This trend is similar to the trend observed inFig. 3(a) for LPAHs, implying the substantial contribution of the LPAHscontent towards the total amount of the 16 PAHs. Conversely, zeolitecatalysts have limited impact on the values of ∑PAH7 which are allbelow 6 μg/g, where a decrease can be even observed. Accordingly, it isfurther confirmed that the addition of zeolite catalysts favors the for-mation of LPAHs most, especially for the yield of naphthalene. On theother hand, TEQs of the bio-oils are relatively stable, retaining at thelevel below 2, which indicates that the toxicity of the bio-oils is prac-tically impervious to the use of catalysts. This observation can be re-sulted from two aspects of consideration. First, although the increase inLPAHs is significant, they are less toxic PAH species (Table 2) that onlyhave a minor contribution to the value of TEQ. Moreover, the con-centrations of MPAHs and HPAHs slightly decreases in the presence ofzeolite catalysts (Fig. 3(b) and (c)), which thereby negatively contributeto the value of TEQ. It should be noted that although most HPAHs arecarcinogenic species with much higher TEQ, the LPAHs and MPAHs areas serious threats to human health as HPAHs, since PAHs with two tofour aromatic hydrocarbon rings, are more potent as co-carcinogens atthe promotional phases of cancer (Ramesh et al., 2004). To this end, thereduction of the amount of the 16 PAHs and the 7 carcinogenic PAHsare of equivalent significance. Clearly, the replacement of N2 with CO2as pyrolysis atmosphere gives rise to a decrease in the values of∑PAH16, ∑PAH7, and TEQ by more than 25% in most cases as shown inTable 4. This additionally indicates that CO2 atmosphere possess theability to inhibit the formation of PAHs no matter whether zeolitecatalysts are applied, consequently reducing the toxicity of the bio-oils.

ZSM-5 zeolite that owns MFI topology is good at producing value-added chemicals such as aromatics in biomass pyrolysis, but it alwayscomes along with PAH byproducts (Lu et al., 2018a; Lu et al., 2018b;Zhang and Zhang, 2017). PAHs formed through biomass pyrolysis arenot only a source of toxicity but also important precursors of coke orcarbonaceous species, which as a result decreases carbon conversionefficiency and causes catalyst deactivation (Stark and Ghoniem, 2017).In current study, the carbon conversion efficiency under different cir-cumstance is evaluated by introducing a factor that equals the ratio ofthe relative area of aromatics against the total concentration of the 16PAHs (ƞc). As shown in Fig. 4(b), the addition of zeolite catalyst gen-erally reduces ƞc into a half, while it is increased by substituting CO2 forN2 as the atmosphere. The maximum ƞc appears at the situation of theaddition of Fe/ZSM-5 zeolite by using CO2 as the pyrolysis atmosphere,corresponding to the value of ƞc that is over 3 times higher with respectto the case of traditional pyrolysis under N2 atmosphere without anycatalyst. In light of this phenomenon, it is reasonable to speculate thatthere is a certain synergetic effect of CO2 and Fe-modified HZSM-5zeolite on biomass pyrolytic conversion.

Unlike feedstocks such as sewage sludge (Hu et al., 2019; Ko et al.,2018) and electronic waste (Lee et al., 2017b; Soler et al., 2018), lig-nocellulosic biomass itself does not originally contain PAHs. As a result,the PAHs that observed in bio-oil can be only formed during the ther-mochemical process in catalytic pyrolysis. For example, a variety ofPAHs are possibly produced from benzene derivatives formed by ar-omatization of olefins and reduction of phenolics, through the Diels-Alder reaction and hydrogen radical abstraction-acetylene additionpathways (Kislov et al., 2005). It is thereby inferred that the presence ofCO2 atmosphere potentially impedes dehydrogenation that occurs in ahomogeneous process, further resulting in an inhibition effect on ad-dition reactions by which PAHs are typically formed. Previous studieshave corroborated that CO2 atmosphere can effectively suppress de-hydrogenation of pyrolysis vapor evolved from the pyrolysis of orangepeel by random bond scissions (Kwon et al., 2019). The authors furtherproposed that the formation of PAHs can be also alleviated by the

reactions between CO2 and volatile organic compounds (Kim et al.,2019a; Lee et al., 2017a; Lee et al., 2017b; Lee et al., 2018). All ex-perimental findings in this study indicate the practical application ofCO2 as an atmosphere during catalytic pyrolysis to enhance the con-version efficiency of thermochemical processes, and the strategicalmeans to suppress PAH species referring as environmental pollutants.

4. Conclusion

Combinations of CO2 atmosphere with zeolites in catalytic pyrolysisof sawdust were evaluated concerning the 16 EPA PAHs. The replace-ment of N2 with CO2 caused a synchronous decrease in LPAHs, MPAHsand HPAHs, and thus abated the ∑PAH16 and ∑PAH7 in bio-oil and itstoxicity. CO2 led to higher carbon conversion efficiency, which reachedthe maximum synergized with Fe/HZSM-5 zeolite, doubled in com-parison with N2. Moreover, furans, ketones and acids decreased, andphenolics and aromatics increased in bio-oil under CO2, exhibiting in-tensified deoxygenation impact. This study experimentally validatedthe promoting effect of CO2 on the quality of bio-oil from catalyticpyrolysis with zeolites.

CRediT authorship contribution statement

Yao He: Conceptualization, Methodology, Writing - original draft.Si Chen: Investigation, Formal analysis. Junjie Chen: Data curation,Visualization. Dongxia Liu: Supervision, Validation. Xunan Ning:Funding acquisition, Resources. Jingyong Liu: Project administration.Tiejun Wang: Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financialinterests or personal relationships that could have appeared to influ-ence the work reported in this paper.

Acknowledgement

This work was financially supported by National Natural ScienceFoundation of China (51806040 and 51978175), 2017 Central SpecialFund for Soil, Preliminary Study on Harmless Treatment andComprehensive Utilization of Tailings in Dabao Mountain (18HK0108),and Key Laboratory of Pollution Control and Ecosystem Restoration inIndustry Clusters, Ministry of Education, South China University ofTechnology, China.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.biortech.2019.122417.

References

Alhroub, I., Kozliak, E., Kubatova, A., Sulkes, M., 2018. PAH/aromatic tar and cokeprecursor formation in the early stages of triglyceride (triolein) pyrolysis. J. Phys.Chem. A 122 (12), 3238–3249.

Bernardon, C., Ben Osman, M., Laugel, G., Louis, B., Pale, P., 2017. Acidity versus metal-induced Lewis acidity in zeolites for Friedel-Crafts acylation. C. R. Chim. 20 (1),20–29.

Che, Q.F., Yang, M.J., Wang, X.H., Chen, X., Chen, W., Yang, Q., Yang, H.P., Chen, H.P.,2019. Aromatics production with metal oxides and ZSM-5 as catalysts in catalyticpyrolysis of wood sawdust. Fuel Process. Technol. 188, 146–152.

Dabros, T.M.H., Stummann, M.Z., Høj, M., Jensen, P.A., Grunwaldt, J.-D., Gabrielsen, J.,Mortensen, P.M., Jensen, A.D., 2018. Transportation fuels from biomass fast pyr-olysis, catalytic hydrodeoxygenation, and catalytic fast hydropyrolysis. Prog. EnergyCombust. Sci. 68, 268–309.

Dandajeh, H.A., Ladommatos, N., Hellier, P., Eveleigh, A., 2018. Influence of carbonnumber of C-1-C-7 hydrocarbons on PAH formation. Fuel 228, 140–151.

Dat, N.D., Chang, M.B., 2017. Review on characteristics of PAHs in atmosphere, an-thropogenic sources and control technologies. Sci. Total Environ. 609, 682–693.

Dunnigan, L., Morton, B.J., van Eyk, P.J., Ashman, P.J., Zhang, X.P., Hall, P.A., Kwong,

Y. He, et al. Bioresource Technology 297 (2020) 122417

7

https://doi.org/10.1016/j.biortech.2019.122417https://doi.org/10.1016/j.biortech.2019.122417http://refhub.elsevier.com/S0960-8524(19)31647-5/h0005http://refhub.elsevier.com/S0960-8524(19)31647-5/h0005http://refhub.elsevier.com/S0960-8524(19)31647-5/h0005http://refhub.elsevier.com/S0960-8524(19)31647-5/h0010http://refhub.elsevier.com/S0960-8524(19)31647-5/h0010http://refhub.elsevier.com/S0960-8524(19)31647-5/h0010http://refhub.elsevier.com/S0960-8524(19)31647-5/h0015http://refhub.elsevier.com/S0960-8524(19)31647-5/h0015http://refhub.elsevier.com/S0960-8524(19)31647-5/h0015http://refhub.elsevier.com/S0960-8524(19)31647-5/h0020http://refhub.elsevier.com/S0960-8524(19)31647-5/h0020http://refhub.elsevier.com/S0960-8524(19)31647-5/h0020http://refhub.elsevier.com/S0960-8524(19)31647-5/h0020http://refhub.elsevier.com/S0960-8524(19)31647-5/h0025http://refhub.elsevier.com/S0960-8524(19)31647-5/h0025http://refhub.elsevier.com/S0960-8524(19)31647-5/h0030http://refhub.elsevier.com/S0960-8524(19)31647-5/h0030http://refhub.elsevier.com/S0960-8524(19)31647-5/h0035

C.W., 2017. Polycyclic aromatic hydrocarbons on particulate matter emitted duringthe co-generation of bioenergy and biochar from rice husk. Bioresour. Technol. 244,1015–1023.

Fabbri, D., Adamiano, A., Torri, C., 2010. GC-MS determination of polycyclic aromatichydrocarbons evolved from pyrolysis of biomass. Anal. Bioanal. Chem. 397 (1),309–317.

Hoff, T.C., Holmes, M.J., Proano-Aviles, J., Emdadi, L., Liu, D.X., Brown, R.C., Tessonnier,J.P., 2017. Decoupling the Role of External Mass Transfer and Intracrystalline PoreDiffusion on the Selectivity of HZSM-5 for the Catalytic Fast Pyrolysis of Biomass.ACS Sustainable Chem. Eng. 5 (10), 8766–8776.

Hu, Y.J., Yu, W.J., Wibowo, H., Xia, Y.Y., Lu, Y.J., Yan, M., 2019. Effect of catalysts ondistribution of polycyclic-aromatic hydrocarbon (PAHs) in bio-oils from the pyrolysisof dewatered sewage sludge at high and low temperatures. Sci. Total Environ. 667,263–270.

Iliopoulou, E.F., Stefanidis, S.D., Kalogiannis, K.G., Delimitis, A., Lappas, A.A.,Triantafyllidis, K.S., 2012. Catalytic upgrading of biomass pyrolysis vapors usingtransition metal-modified ZSM-5 zeolite. Appl. Catal. B-Environ. 127, 281–290.

Kerr, J.M., Melton, H.R., McMillen, S.J., Magaw, R.I., Naughton, G. 1999. PolyaromaticHydrocarbon Content in Crude Oils Around the World. in: SPE/EPA Exploration andProduction Environmental Conference, 1-3 March, Society of Petroleum Engineers.Austin, Texas.

Kim, J.H., Oh, J.I., Lee, J., Kwon, E.E., 2019a. Valorization of sewage sludge via a pyr-olytic platform using carbon dioxide as a reactive gas medium. Energy 179, 163–172.

Kim, Y., Lee, J., Yi, H., Tsang, Y.F., Kwon, E.E., 2019b. Investigation into role of CO2 intwo-stage pyrolysis of spent coffee grounds. Bioresour. Technol. 272, 48–53.

Kislov, V.V., Islamova, N.I., Kolker, A.M., Lin, S.H., Mebel, A.M., 2005. Hydrogen ab-straction acetylene addition and diels−alder mechanisms of PAH Formation: a de-tailed study using first principles calculations. J. Chem. Theory Comput. 1 (5),908–924.

Ko, J.H., Wang, J.C., Xu, Q.Y., 2018. Impact of pyrolysis conditions on polycyclic aro-matic hydrocarbons (PAHs) formation in particulate matter (PM) during sewagesludge pyrolysis. Chemosphere 208, 108–116.

Konczak, M., Gao, Y.Z., Oleszczuk, P., 2019. Carbon dioxide as a carrier gas and biomassaddition decrease the total and bioavailable polycyclic aromatic hydrocarbons inbiochar produced from sewage sludge. Chemosphere 228, 26–34.

Kumar, A., Kumar, A., Kumar, J., Bhaskar, T., 2019. Catalytic pyrolysis of soda lignin overzeolites using pyrolysis gas chromatography-mass spectrometry. Bioresour. Technol.291.

Kwon, D., Oh, J.I., Lambe, S.S., Moon, D.H., Kwon, E.E., 2019. Orange peel valorizationby pyrolysis under the carbon dioxide environment. Bioresour. Technol. 285.

Kwon, E.E., Yi, H., Castaldi, M.J., 2012. Utilizing carbon dioxide as a reaction medium tomitigate production of polycyclic aromatic hydrocarbons from the thermal decom-position of styrene butadiene rubber. Environ. Sci. Technol. 46 (19), 10752–10757.

Lee, J., Choi, D., Tsang, Y.F., Oh, J.-I., Kwon, E.E., 2017a. Employing CO2 as reactionmedium for in-situ suppression of the formation of benzene derivatives and poly-cyclic aromatic hydrocarbons during pyrolysis of simulated municipal solid waste.Environ. Pollut. 224, 476–483.

Lee, J., Lee, T., Ok, Y.S., Oh, J.-I., Kwon, E.E., 2017b. Using CO2 to mitigate evolution ofharmful chemical compounds during thermal degradation of printed circuit boards. J.CO2 Util. 20, 66–72.

Lee, T., Oh, J.I., Kim, T., Tsang, D.C.W., Kim, K.H., Lee, J., Kwon, E.E., 2018. Controllinggeneration of benzenes and polycyclic aromatic hydrocarbons in thermolysis ofpolyvinyl chloride in CO2. Energy Convers. Manage. 164, 453–459.

Li, P., Chen, X., Wang, X.H., Shao, J.G., Lin, G.Y., Yang, H.P., Yang, Q., Chen, H.P., 2017.Catalytic Upgrading of Fast Pyrolysis Products with Fe-, Zr-, and Co-Modified ZeolitesBased on Pyrolyzer-GC/MS Analysis. Energy Fuels 31 (4), 3979–3986.

Li, P., Li, D., Yang, H.P., Wang, X.H., Chen, H.P., 2016. Effects of Fe-, Zr-, and Co-Modified Zeolites and Pretreatments on Catalytic Upgrading of Biomass Fast PyrolysisVapors. Energy Fuels 30 (4), 3004–3013.

Lu, Q., Guo, H.Q., Zhou, M.X., Cui, M.S., Dong, C.Q., Yang, Y.P., 2018a. Selective pre-paration of monocyclic aromatic hydrocarbons from catalytic cracking of biomassfast pyrolysis vapors over Mo2N/HZSM-5 catalyst. Fuel Process. Technol. 173,134–142.

Lu, Q., Guo, H.Q., Zhou, M.X., Zhang, Z.X., Cui, M.S., Zhang, Y.Y., Yang, Y.P., Zhang, L.B.,2018b. Monocyclic aromatic hydrocarbons production fromcatalytic cracking of pinewood-derived pyrolytic vapors over Ce-Mo2N/HZSM-5 catalyst. Sci. Total Environ.634, 141–149.

Mahler, B.J., Van Metre, P.C., Crane, J.L., Watts, A.W., Scoggins, M., Williams, E.S., 2012.Coal-Tar-Based Pavement Sealcoat and PAHs: Implications for the Environment,Human Health, and Stormwater Management. Environ. Sci. Technol. 46 (6),3039–3045.

Maliutina, K., Tahmasebi, A., Yu, J.L., Saltykov, S.N., 2017. Comparative study on flashpyrolysis characteristics of microalgal and lignocellulosic biomass in entrained-flowreactor. Energy Convers. Manage. 151, 426–438.

Mante, O.D., Agblevor, F.A., Oyama, S.T., McClung, R., 2012. The influence of recyclingnon-condensable gases in the fractional catalytic pyrolysis of biomass. Bioresour.Technol. 111, 482–490.

Mohabeer, C., Reyes, L., Abdelouahed, L., Marcotte, S., Taouk, B., 2019. Investigatingcatalytic de-oxygenation of cellulose, xylan and lignin bio-oils using HZSM-5 and Fe-HZSM-5. J. Anal. Appl. Pyrol. 137, 118–127.

Mullen, C.A., Boateng, A.A., 2015. Production of Aromatic Hydrocarbons via CatalyticPyrolysis of Biomass over Fe-Modified HZSM-5 Zeolites. ACS Sustainable Chem. Eng.3 (7), 1623–1631.

Nguyen, V.T., Nguyen, T.B., Chen, C.W., Hung, C.M., Vo, T.D.H., Chang, J.H., Dong, C.D.,2019. Influence of pyrolysis temperature on polycyclic aromatic hydrocarbons pro-duction and tetracycline adsorption behavior of biochar derived from spent coffeeground. Bioresour. Technol. 284, 197–203.

Nisbet, I.C.T., Lagoy, P.K., 1992. Toxic Equivalency Factors (Tefs) for PolycyclicAromatic-Hydrocarbons (Pahs). Regul. Toxicol. Pharm. 16 (3), 290–300.

Odabasi, M., Cetin, E., Sofuoglu, A., 2006. Determination of octanol-air partition coeffi-cients and supercooled liquid vapor pressures of PAHs as a function of temperature:application to gas-particle partitioning in an urban atmosphere. Atmos. Environ. 40(34), 6615–6625.

Pilon, G., Lavoie, J.M., 2013. Pyrolysis of switchgrass (Panicum virgatum L.) at lowtemperatures in N-2 and CO2 environments; a study on chemical composition ofchars extracts and bio-oils. J. Anal. Appl. Pyrol. 101, 122–131.

Ramesh, A., Walker, S.A., Hood, D.B., Guillen, M.D., Schneider, K., Weyand, E.H., 2004.Bioavailability and risk assessment of orally ingested polycyclic aromatic hydro-carbons. Int. J. Toxicol. 23 (5), 301–333.

Sharifzadeh, M., Sadeqzadeh, M., Guo, M., Borhani, T.N., Murthy Konda, N.V.S.N.,Garcia, M.C., Wang, L., Hallett, J., Shah, N., 2019. The multi-scale challenges ofbiomass fast pyrolysis and bio-oil upgrading: Review of the state of art and futureresearch directions. Prog. Energy Combust. Sci. 71, 1–80.

Soler, A., Conesa, J.A., Iniguez, M.E., Ortuno, N., 2018. Pollutant formation in the pyr-olysis and combustion of materials combining biomass and e-waste. Sci. TotalEnviron. 622, 1258–1264.

Stark, A.K., Ghoniem, A.F., 2017. Quantification of the influence of particle diameter onPolycyclic Aromatic Hydrocarbon (PAH) formation in fluidized bed biomass pyr-olysis. Fuel 206, 276–288.

Szegedi, A., Konya, Z., Mehn, D., Solymar, E., Pal-Borbely, G., Horvath, Z.E., Biro, L.P.,Kiricsi, I., 2004. Spherical mesoporous MCM-41 materials containing transition me-tals: synthesis and characterization. Appl. Catal. A-General 272 (1–2), 257–266.

Wang, C.Y., Wang, Y.D., Herath, H., 2017. Polycyclic aromatic hydrocarbons (PAHs) inbiochar - Their formation, occurrence and analysis: a review. Org Geochem. 114,1–11.

Wang, M., Dewil, R., Maniatis, K., Wheeldon, J., Tan, T., Baeyens, J., Fang, Y., 2019.Biomass-derived aviation fuels: Challenges and perspective. Prog. Energy Combust.Sci. 74, 31–49.

Xu, C.X., Jiang, B.B., Liao, Z.W., Wang, J.D., Huang, Z.L., Yang, Y.R., 2017. Effect of metalon the methanol to aromatics conversion over modified ZSM-5 in the presence ofcarbon dioxide. RSC Adv. 7 (18), 10729–10736.

Zhang, B., Zhang, J., 2017. Influence of Reaction Atmosphere (N-2, CO, CO2, and H-2) onZSM-5 Catalyzed Microwave-Induced Fast Pyrolysis of Medicinal Herb Residue forBiofuel Production. Energy Fuels 31 (9), 9627–9632.

Zhang, H.Y., Xiao, R., Wang, D.H., He, G.Y., Shao, S.S., Zhang, J.B., Zhong, Z.P., 2011.Biomass fast pyrolysis in a fluidized bed reactor under N-2, CO2, CO, CH4 and H-2atmospheres. Bioresour. Technol. 102 (5), 4258–4264.

Zhou, H., Wu, C.F., Meng, A.H., Zhang, Y.G., Williams, P.T., 2014a. Effect of interactionsof biomass constituents on polycyclic aromatic hydrocarbons (PAH) formation duringfast pyrolysis. J. Anal. Appl. Pyrol. 110, 264–269.

Zhou, H., Wu, C.F., Onwudili, J.A., Meng, A.H., Zhang, Y.G., Williams, P.T., 2014b.Polycyclic Aromatic Hydrocarbon Formation from the Pyrolysis/Gasification ofLignin at Different Reaction Conditions. Energy Fuels 28 (10), 6371–6379.

Y. He, et al. Bioresource Technology 297 (2020) 122417

8

http://refhub.elsevier.com/S0960-8524(19)31647-5/h0035http://refhub.elsevier.com/S0960-8524(19)31647-5/h0035http://refhub.elsevier.com/S0960-8524(19)31647-5/h0035http://refhub.elsevier.com/S0960-8524(19)31647-5/h0040http://refhub.elsevier.com/S0960-8524(19)31647-5/h0040http://refhub.elsevier.com/S0960-8524(19)31647-5/h0040http://refhub.elsevier.com/S0960-8524(19)31647-5/h0045http://refhub.elsevier.com/S0960-8524(19)31647-5/h0045http://refhub.elsevier.com/S0960-8524(19)31647-5/h0045http://refhub.elsevier.com/S0960-8524(19)31647-5/h0045http://refhub.elsevier.com/S0960-8524(19)31647-5/h0050http://refhub.elsevier.com/S0960-8524(19)31647-5/h0050http://refhub.elsevier.com/S0960-8524(19)31647-5/h0050http://refhub.elsevier.com/S0960-8524(19)31647-5/h0050http://refhub.elsevier.com/S0960-8524(19)31647-5/h0055http://refhub.elsevier.com/S0960-8524(19)31647-5/h0055http://refhub.elsevier.com/S0960-8524(19)31647-5/h0055http://refhub.elsevier.com/S0960-8524(19)31647-5/h0065http://refhub.elsevier.com/S0960-8524(19)31647-5/h0065http://refhub.elsevier.com/S0960-8524(19)31647-5/h0070http://refhub.elsevier.com/S0960-8524(19)31647-5/h0070http://refhub.elsevier.com/S0960-8524(19)31647-5/h0075http://refhub.elsevier.com/S0960-8524(19)31647-5/h0075http://refhub.elsevier.com/S0960-8524(19)31647-5/h0075http://refhub.elsevier.com/S0960-8524(19)31647-5/h0075http://refhub.elsevier.com/S0960-8524(19)31647-5/h0080http://refhub.elsevier.com/S0960-8524(19)31647-5/h0080http://refhub.elsevier.com/S0960-8524(19)31647-5/h0080http://refhub.elsevier.com/S0960-8524(19)31647-5/h0085http://refhub.elsevier.com/S0960-8524(19)31647-5/h0085http://refhub.elsevier.com/S0960-8524(19)31647-5/h0085http://refhub.elsevier.com/S0960-8524(19)31647-5/h0090http://refhub.elsevier.com/S0960-8524(19)31647-5/h0090http://refhub.elsevier.com/S0960-8524(19)31647-5/h0090http://refhub.elsevier.com/S0960-8524(19)31647-5/h0095http://refhub.elsevier.com/S0960-8524(19)31647-5/h0095http://refhub.elsevier.com/S0960-8524(19)31647-5/h0100http://refhub.elsevier.com/S0960-8524(19)31647-5/h0100http://refhub.elsevier.com/S0960-8524(19)31647-5/h0100http://refhub.elsevier.com/S0960-8524(19)31647-5/h0105http://refhub.elsevier.com/S0960-8524(19)31647-5/h0105http://refhub.elsevier.com/S0960-8524(19)31647-5/h0105http://refhub.elsevier.com/S0960-8524(19)31647-5/h0105http://refhub.elsevier.com/S0960-8524(19)31647-5/h0110http://refhub.elsevier.com/S0960-8524(19)31647-5/h0110http://refhub.elsevier.com/S0960-8524(19)31647-5/h0110http://refhub.elsevier.com/S0960-8524(19)31647-5/h0115http://refhub.elsevier.com/S0960-8524(19)31647-5/h0115http://refhub.elsevier.com/S0960-8524(19)31647-5/h0115http://refhub.elsevier.com/S0960-8524(19)31647-5/h0120http://refhub.elsevier.com/S0960-8524(19)31647-5/h0120http://refhub.elsevier.com/S0960-8524(19)31647-5/h0120http://refhub.elsevier.com/S0960-8524(19)31647-5/h0125http://refhub.elsevier.com/S0960-8524(19)31647-5/h0125http://refhub.elsevier.com/S0960-8524(19)31647-5/h0125http://refhub.elsevier.com/S0960-8524(19)31647-5/h0130http://refhub.elsevier.com/S0960-8524(19)31647-5/h0130http://refhub.elsevier.com/S0960-8524(19)31647-5/h0130http://refhub.elsevier.com/S0960-8524(19)31647-5/h0130http://refhub.elsevier.com/S0960-8524(19)31647-5/h0135http://refhub.elsevier.com/S0960-8524(19)31647-5/h0135http://refhub.elsevier.com/S0960-8524(19)31647-5/h0135http://refhub.elsevier.com/S0960-8524(19)31647-5/h0135http://refhub.elsevier.com/S0960-8524(19)31647-5/h0140http://refhub.elsevier.com/S0960-8524(19)31647-5/h0140http://refhub.elsevier.com/S0960-8524(19)31647-5/h0140http://refhub.elsevier.com/S0960-8524(19)31647-5/h0140http://refhub.elsevier.com/S0960-8524(19)31647-5/h0145http://refhub.elsevier.com/S0960-8524(19)31647-5/h0145http://refhub.elsevier.com/S0960-8524(19)31647-5/h0145http://refhub.elsevier.com/S0960-8524(19)31647-5/h0150http://refhub.elsevier.com/S0960-8524(19)31647-5/h0150http://refhub.elsevier.com/S0960-8524(19)31647-5/h0150http://refhub.elsevier.com/S0960-8524(19)31647-5/h0155http://refhub.elsevier.com/S0960-8524(19)31647-5/h0155http://refhub.elsevier.com/S0960-8524(19)31647-5/h0155http://refhub.elsevier.com/S0960-8524(19)31647-5/h0160http://refhub.elsevier.com/S0960-8524(19)31647-5/h0160http://refhub.elsevier.com/S0960-8524(19)31647-5/h0160http://refhub.elsevier.com/S0960-8524(19)31647-5/h0165http://refhub.elsevier.com/S0960-8524(19)31647-5/h0165http://refhub.elsevier.com/S0960-8524(19)31647-5/h0165http://refhub.elsevier.com/S0960-8524(19)31647-5/h0165http://refhub.elsevier.com/S0960-8524(19)31647-5/h0170http://refhub.elsevier.com/S0960-8524(19)31647-5/h0170http://refhub.elsevier.com/S0960-8524(19)31647-5/h0175http://refhub.elsevier.com/S0960-8524(19)31647-5/h0175http://refhub.elsevier.com/S0960-8524(19)31647-5/h0175http://refhub.elsevier.com/S0960-8524(19)31647-5/h0175http://refhub.elsevier.com/S0960-8524(19)31647-5/h0180http://refhub.elsevier.com/S0960-8524(19)31647-5/h0180http://refhub.elsevier.com/S0960-8524(19)31647-5/h0180http://refhub.elsevier.com/S0960-8524(19)31647-5/h0185http://refhub.elsevier.com/S0960-8524(19)31647-5/h0185http://refhub.elsevier.com/S0960-8524(19)31647-5/h0185http://refhub.elsevier.com/S0960-8524(19)31647-5/h0190http://refhub.elsevier.com/S0960-8524(19)31647-5/h0190http://refhub.elsevier.com/S0960-8524(19)31647-5/h0190http://refhub.elsevier.com/S0960-8524(19)31647-5/h0190http://refhub.elsevier.com/S0960-8524(19)31647-5/h0195http://refhub.elsevier.com/S0960-8524(19)31647-5/h0195http://refhub.elsevier.com/S0960-8524(19)31647-5/h0195http://refhub.elsevier.com/S0960-8524(19)31647-5/h0200http://refhub.elsevier.com/S0960-8524(19)31647-5/h0200http://refhub.elsevier.com/S0960-8524(19)31647-5/h0200http://refhub.elsevier.com/S0960-8524(19)31647-5/h0205http://refhub.elsevier.com/S0960-8524(19)31647-5/h0205http://refhub.elsevier.com/S0960-8524(19)31647-5/h0205http://refhub.elsevier.com/S0960-8524(19)31647-5/h0210http://refhub.elsevier.com/S0960-8524(19)31647-5/h0210http://refhub.elsevier.com/S0960-8524(19)31647-5/h0210http://refhub.elsevier.com/S0960-8524(19)31647-5/h0215http://refhub.elsevier.com/S0960-8524(19)31647-5/h0215http://refhub.elsevier.com/S0960-8524(19)31647-5/h0215http://refhub.elsevier.com/S0960-8524(19)31647-5/h0220http://refhub.elsevier.com/S0960-8524(19)31647-5/h0220http://refhub.elsevier.com/S0960-8524(19)31647-5/h0220http://refhub.elsevier.com/S0960-8524(19)31647-5/h0225http://refhub.elsevier.com/S0960-8524(19)31647-5/h0225http://refhub.elsevier.com/S0960-8524(19)31647-5/h0225http://refhub.elsevier.com/S0960-8524(19)31647-5/h0230http://refhub.elsevier.com/S0960-8524(19)31647-5/h0230http://refhub.elsevier.com/S0960-8524(19)31647-5/h0230http://refhub.elsevier.com/S0960-8524(19)31647-5/h0235http://refhub.elsevier.com/S0960-8524(19)31647-5/h0235http://refhub.elsevier.com/S0960-8524(19)31647-5/h0235http://refhub.elsevier.com/S0960-8524(19)31647-5/h0240http://refhub.elsevier.com/S0960-8524(19)31647-5/h0240http://refhub.elsevier.com/S0960-8524(19)31647-5/h0240

Consequence of replacing nitrogen with carbon dioxide as atmosphere on suppressing the formation of polycyclic aromatic hydrocarbons in catalytic pyrolysis of sawdustIntroductionMaterials and methodMaterials and catalyst preparationCatalyst characterizationPyrolysis experimentBio-oil pretreatment and analysis

Results and discussionCatalyst characterizationComposition of bio-oilsCO2 suppression for PAHsPAHs determinationCO2 effect on PAHs with different molecular weightsCO2 effect on toxicity and carbon conversion efficiency

ConclusionCRediT authorship contribution statementmk:H1_16AcknowledgementSupplementary dataReferences